Over ons 🤖

Laten we elkaar leren kennen

Vertel me de missie en visie

Leg het verhaal achter Mach8 uit

Hallo daar 👋

Hoe kunnen we je helpen?

Mijn gegevens mogen worden gebruikt om me op de hoogte te houden van relevant nieuws van Mach8

E-commerce & AI·7 min·4 May 2025

AI and e-commerce personalisation: what is already possible today?

Personalisation in e-commerce is often presented as the future. But much of what is described as a future prospect is already reality. This is an honest overview of what is possible today and where the limits lie.

Any webshop with sufficient customer data can personalise today. The question is not whether it is possible, but which forms of personalisation are worthwhile and what the real technical threshold is to get started.

Product recommendations: a reality for years

The most mature form of AI personalisation in e-commerce is product recommendations. Algorithms that analyse what customers buy, view, and compare, then recommend related products on that basis.

This is not new technology. Amazon built its recommendation system in the early 2000s. What has changed is the cost and accessibility: cloud platforms such as AWS Personalize, Google Recommendations AI, and Coveo offer this as a service you can integrate without building your own data science team.

For mid-sized webshops the thresholds are real: you need sufficient transaction data (roughly 1,000+ transactions per month for stable recommendations), and a platform that supports the integration.

Dynamic content: the same, but different

A step beyond product recommendations is dynamic content: the homepage, category page, or even product page that differs per visitor based on behaviour, location, or purchase history.

Tools such as Nosto, Dynamic Yield, and Salesforce Personalization make this possible without custom-built infrastructure. You define segments, link them to content variants, and the platform handles the display.

The pitfall is segmentation granularity: segments that are too broad produce no real personalisation, segments that are too narrow have insufficient data for reliable insights. Finding the right middle ground requires experimentation and iteration.

Email personalisation: mature and proven

Personalised email is one of the most proven applications of AI in e-commerce. Abandoned cart emails, reactivation campaigns, product recommendations based on browse history: this works and the technology is mature.

The ROI is well documented: personalised emails have significantly higher open rates and click-through rates than generic newsletters. The investment lies in setting up the data flows and defining the trigger logic, not in complex AI development.

Chatbots as a personal advice channel

A chatbot that uses customer data can provide a limited form of personalised advice: based on previously purchased products, open orders, or stated preferences.

This is not a human advice conversation, but it is contextually more relevant than a generic FAQ. The limitations are clear: the chatbot only knows what is in its systems and cannot pick up nuances that a human agent would notice. But for routine queries with known customers it works well.

See also our page on chatbots for more on what Mach8 builds in this area.

Search personalisation: underestimated potential

Search personalisation is one of the most underestimated forms of AI in e-commerce. Adapting the search results on your webshop to the preferences and history of the logged-in visitor has a direct effect on conversion.

If a customer always buys in the budget segment, searching for "laptop" shows laptops in that segment first. If a customer previously bought a particular brand, that brand rises in search results. This is not radical technology, but it requires that your search platform supports it and that you consciously configure the relevance parameters.

What does not work well yet

Being honest means naming the limits too:

  • Hyper-personalisation at page level for unknown visitors does not work without data. A first visit is always generic.
  • Real-time personalisation based on intent requires fast inference and is technically complex and costly for most webshops.
  • Personalisation that genuinely feels bespoke requires significant customer data over a longer period. A visitor with two purchases has an insufficient profile for deep personalisation.

Conclusion

AI personalisation in e-commerce is widely applicable today: product recommendations, dynamic content, personalised email, and search optimisation are not a future vision but available technology. The key is matching the technical complexity to the available data and resources.

Mach8 helps with selecting and implementing the right personalisation approach for your e-commerce platform. Get in touch for an analysis of your situation.

Ready to apply AI?

We help you go from strategy to implementation. Schedule a no-obligation call.

Schedule a call